The present invention relates generally to systems for transmitting data over fiber optic networks, and specifically to systems for transmitting an encoded Internet protocol (IP) packet directly over an optical or photonic layer of a fiber optic network without the overhead and expense of adapting the IP packet to another encoding format.
The global Internet has become a gateway to information for the general public. Given the online availability of this seemingly endless resource of information, a desire to quickly access this information has driven an evolution in the telecommunications industry and, in particular, in telecommunications networks. As telecommunications networks evolve, the growth in data traffic on the networks appears to outpace the growth in voice traffic. Some industry commentators have noted that the average packet size in telecommunications networks has increased from 280 bytes to 330 bytes in 1998 alone. Additionally, these commentators have noted that the average session duration has decreased down to only 13 seconds. In summary, the growth in data traffic along with changing characteristics of the data traffic itself has forced many telecommunications service providers to seek a more efficient and higher bandwidth mechanism for transporting data traffic.
One efficient type of telecommunications network for transporting such data traffic is a packet-based telecommunications network. A packet-based telecommunications network is a communication network using protocols in which messages are divided into packets before they are sent. Each packet is individually transmitted and can follow different routes to its destination within the network. Once all of the packets forming the message have arrived at the destination, the message is recompiled from the packets and delivered as a whole message.
An example of such a packet-based telecommunications network is an IP-based network. Essentially, an IP-based network uses Internet protocol or IP as the protocol to handle addressing of packets. Some networks combine IP with a higher level protocol called transport control protocol (TCP), which establishes a virtual connection or link between a destination and a source. Those skilled in the art will be familiar with IP-based networks and the current version of IP called IPv4, which is promulgated by the Internet Engineering Task Force (IETF). A newer version of IP called IPv6 or IPng (IP Next Generation) is currently under development by IETF.
Many data networks generally rely upon a conventional Synchronous Optical Network (SONET) layer to transport data within the network. This is due to the ability of SONET to support international standards for interconnecting networks and SONET's desirably efficient use of bandwidth. Basically, SONET defines interface standards beginning at the physical layer level of the Open Systems Interconnection (OSI) Reference Network Model developed by the International Organization for Standardization (ISO). The interface standards defined by SONET describe a synchronous hierarchy of interface rates that allow data sequences at differing rates to be multiplexed within a SONET frame, which is then transmitted as an optical signal. The synchronous hierarchy of SONET results in an overhead percentage that does not vary as the rate increases. Each time one multiplexes to a higher rate, the lower-rate signals are synchronously byte-interleaved to produce the higher rate. Since the lower-rate signals are synchronous to one another, no additional overhead is needed (e.g., with stuffing bytes and associated signaling) to support rate matching. Those skilled in the art will be familiar with the OSI Reference Network Model and additional features of SONET standards, equipment and systems.
Furthermore, those skilled in the art will understand that conventional SONET systems are implemented with multiple layers. In general, each of the layers are responsible for specific functions. The lowest layer is a photonic or optical layer while other layers add functionality (such as framing, scrambling, and error monitoring) used by devices in the network. The photonic or optical layer essentially converts electrical signals to optical signals and is responsible for maintaining the pulse shape, wavelength and power levels of the optical signals. The photonic or optical layer is analogous to the physical layer as described in the OSI Reference Network Model.
While each of the other layers in SONET systems has a particular amount of overhead associated the particular layer's functionality added to the system, the photonic or optical layer has none. In other words, at the bottom layer of the network model, the photonic or optical layer adds no extra processing or data overhead to a packet of information when transmitting an optical signal representing the packet directly over the photonic or optical layer of the network.
Due to the synchronous nature of SONET, one problem that may be encountered when using SONET to transport IP packets is maintaining a synchronous clock within the network at all times. Applicant has observed that long pattern sequences of the same value may undesirably cause loss of a clock signal in the network. Typically, the bits of a packet transmitted directly over the photonic or optical layer are encoded in a conventional non-return-to-zero (NRZ) encoding format. After a finite number of consecutive bits at the same logic level, the clock signal may drop out because the optical signal being transmitted remains at a constant, DC, unmodulated level as a quasi-continuous wave state. Once the clock signal drops out, future packets may be lost when they arrive at the optical receiver because the timing of the packet is no longer synchronous.
In order to avoid this problem and help maintain a synchronous clock, the SONET layer for transporting IP packets usually scrambles the payload (conventionally referred to as a Synchronous Payload Element) of the packet. Scrambling of the payload essentially breaks up long pattern sequences that may be input to SONET equipment within the network. By breaking up the long pattern sequences, the electro-optics within SONET equipment, such as clock recovery circuitry in optical receivers, can function properly and packets are not lost.
Unfortunately, extra hardware and processing is required to scramble and unscramble (i.e., multiplex and demultiplex) the payload of an IP packet in SONET. In other words, another layer of encoding or adapting of the IP packet is required in addition to encoding bits of the IP packet directly into light pulses. For example, a conventional statistical time division multiplexer is a device within a router that accepts electrical input from a variety of sources and provides a multiplexed optical signal output representing a SONET frame (which may include an IP packet). The statistical multiplexer also has scrambling hardware to enable the router to ensure that a clock signal can be easily recovered by the receiver of the SONET frame. This extra hardware and the associated overhead for scrambling the payload of each packet is costly in terms of monetary costs, wasted bandwidth and processing resources within the network. Thus, Applicant has discovered that it would be advantageous to transmit IP packets directly over the optical or photonic layer without having to adapt the IP packet any further while simultaneously maintaining the ability to extract a clock signal at all times.
Point-to-Point (PPP) protocol is a standard method of transmitting different protocol packets, including IP packets, over point-to-point links in the network. Thus, an IP packet may be framed for transmission in the network using PPP in a frame, such as an High Level Data Link Control (HDLC)-like frame. Essentially, HDLC-like frames conform to a protocol at a data link layer allowing for control data flow and error correction. PPP in HDLC-like frames allows for framing of both bit-oriented and octet-oriented synchronous and asynchronous links. Those skilled in the art will be familiar with PPP in HDLC-like frames. Additional information about PPP in HDLC-like frames is available in RFC 1662 entitled "PPP in HDLC-like Framing" promulgated and distributed by IETF.
Another problem associated with transmitting such IP packets within a fiber optic network is that when no IP packets are being transmitted from an IP router, the IP router is in an idle state. In the idle state, the IP router must transmit some type of signal in order to maintain a synchronous clock at the receiver. Thus, the IP router typically transmits a filler signal during the idle state. Using PPP in an HDLC-like frame, the filler signal is usually characterized as a series of continuous logical ones encoded in the NRZ encoding format. Applicant has observed that in such a situation and without adapting the filler signal further, the receiver can easily lose the clock signal after a certain number of logical ones are transmitted on the photonic or optical layer.
This problem is not apparent in networks transmitting data using conventional Asynchronous Transfer Mode (ATM) transmissions because the filler has a short pattern sequence of logical ones and zeros, such as "01101010", which is repeated 48 times within the filler signal's payload. In this manner, a clock signal can be recovered when using ATM for transmitting packets directly over an optical layer. However, using ATM requires an undesirably large transport overhead and is not as efficient as IP packet transmission. For example, some lightwave industry commentators have published articles describing the ATM transport overhead as a major deficiency in network bandwidth efficiency because the ATM transport overhead can be up to 50% of the packet payload depending upon the size of the packet.
Patents and publications have generally described extracting clock signals from optical signals encoded in both NRZ encoding format and RZ encoding format. For example, U.S. Pat. No. 5,339,185 discloses an optical timing extraction circuit. The '185 patent discloses the optical timing extraction circuit capable of extracting an optical clock signal of greater than tens of Gbits/sec from an NRZ encoded signal.
Additionally, U.S. Pat. No. 5,602,862 discloses extracting a clock signal from an optical input signal encoded in either NRZ encoding format or RZ encoding format. The '862 patent discloses a transmitter generating a 5 Gbits/sec optical data signal encoded in an RZ encoding format that is then multiplexed into a 20 Gbits/sec optical pulse stream. The data signal is transmitted to a receiver where the optical data signal is received and a clock signal is optically recovered as opposed to converting the optical data signal to an electrical signal prior to clock recovery. Additionally, the '862 patent discloses operation of the optical clock extraction in the context of injecting the 5 Gbit/sec RZ encoded optical data signal directly into a novel optical clock recovery circuit.